JOURNAL OF VIROLOGY, Mar. 1992, p. 1799-18030022-538X/92/031799-05$02.00/0Copyright ©) 1992, American Society for Microbiology

Mutational Analysis of Conserved N-Linked Glycosylation Sites ofHuman Immunodeficiency Virus Type 1 gp4l

WOAN-RUOH LEE, XIAO-FANG YU, WAN-JR SYU,t MAX ESSEX, AND TUN-HOU LEE*

Department of Cancer Biology, Harvard University School of Public Health, Boston, Massachusetts 02115

Received 10 July 1991/Accepted 4 December 1991

Amino acid substitutions were introduced into four conserved N-linked glycosylation sites of the humanimmunodeficiency virus type 1 envelope transmembrane glycoprotein, gp4l, to alter the canonical N-linkedglycosylation sequences. One altered site produced a severe impairment of viral infectivity, which raises thepossibility that N-linked sugars at this site may have an important role in the human immunodeficiency virustype 1 life cycle.

The molecule that provides the membrane anchor for thehuman immunodeficiency virus type 1 (HIV-1) externalenvelope protein, gpl20 (1), is the envelope transmembraneprotein, gp4l (22). The coding sequence of gp4l is unusuallylong compared with the transmembrane protein of retrovi-ruses outside the subfamily of lentiviruses (14, 17, 18, 21,23). The extra coding sequence of gp4l is located distal to a

characteristic hydrophobic membrane anchor domain and isbelieved to be on the cytoplasmic side of the membrane (2).Except for this unique feature, gp41 has several structuralfeatures, within the region proximal to the hydrophobicmembrane anchor domain, in common with the envelopetransmembrane proteins of a divergent family of animalretroviruses (5, 21). These common features, from the Nterminus to the C terminus of gp4l, include the following: a

hydrophobic stretch of amino acid residues, a region rich inthreonine and serine residues, a string of amino acid residueswith a high probability of forming an a-helix structure, twoor three vicinal cysteine residues, and consensus N-linkedglycosylation sites which are followed by the hydrophobicmembrane anchor domain.The functional significance of some of the conserved

structural features of gp4l was explored in several previousstudies. For instance, mutations introduced to the hydropho-bic sequences in the N terminus of gp4l were found togreatly reduce the syncytium-forming abilities of the mutantviruses (8). Mutagenesis through insertion of in-frame link-ers to the region rich in threonine and serine residues was

found to disrupt the association between gp4l and gpl20 (9).In addition, substituting the two highly conserved vicinalcysteine residues of gp4l with other amino acid residues wasfound to affect the processing of the gpl60 precursor, whichsuggests that the conformation dictated by the disulfide bondformed by these vicinal cysteine residues is important for thematuration of the envelope protein (20).

In the coding sequence for gp41 of most HIV-1 isolateswith known nucleotide sequences, there are four consensus

N-linked glycosylation sites located within a region flankedby two highly conserved vicinal cysteine residues and a

hydrophobic membrane anchor domain (15). This high de-gree of structural conservation is intriguing in light of thesequence heterogeneity found in the envelope genes of

* Corresponding author.t Present address: Graduate Institute of Microbiology and Immu-

nology, National Yang-Ming Medical College, Taipei, Taiwan.

various HIV-1 isolates (15). In the present study, we used an

HIV-1 molecular clone, HXB2 (4), as a working model toaddress the question of whether alteration of these highlyconserved N-linked glycosylation sites of gp4l affects HIV-1infectivity.As shown in Fig. 1, five potential N-linked glycosylation

sites are present within the region flanked by the two highlyconserved vicinal cysteine residues and the hydrophobicmembrane anchor domain of HXB2. Among these five sites,only the first four sites are shared by most HIV-1 isolates(15). Five N-linked glycosylation site mutants of HXB2,designated 611, 616, 624, 637, and 674 (Fig. 1), which had theasparagine residue of the canonical N-linked glycosylationsequence replaced by a noncanonical residue, were con-

structed to study whether these conserved N-linked se-

quences were critical for HIV-1 infectivity. These mutantswere generated by the oligonucleotide-directed mutagenesisof Kunkel et al. (10). The template used was a 2.7-kbSalI-BamHI fragment of HXB2, which contains most of theenvelope-coding sequence of HXB2. The mutagenic oligo-nucleotides used to generate these mutants are shown inTable 1. As shown in Fig. 1, the asparagine residue of thecanonical Asn-X-Ser/Thr sequence was replaced by a histi-dine residue. The SalI-BamHI fragments containing thedesired mutations were verified by DNA sequencing (19),and the M13 replicative forms containing the desired muta-tions were used to replace the corresponding SalI-BamHIfragment of the wild-type virus.Western blot (immunoblot) analyses were first carried out

to examine whether alterations introduced into these con-

served N-linked sites affected the expression of the envelopegene-encoded gp160 and gpl20 (3). For this purpose, eachproviral DNA of the wild-type and mutant viruses was

transfected into COS-7 cells by the DEAE-dextran method(25). Cell lysates were prepared 48 h posttransfection andanalyzed with a reference serum from an HIV-1-seropositivepatient. As shown in Fig. 1, gp160 and gp120 were detectedin the cells transfected with each of the five N-linkedglycosylation site mutants and the wild-type virus. Thespecificities of gpl60 and gp120 were further indicated by theabsence of these two species in the mock-transfected COS-7cells and by the reactivities of these two species to a

hyperimmune sheep antiserum to gpl20 (AIDS Research andReference Reagent Program no. 288; data not shown). Theseresults indicated that the expression of the envelope precur-sor, gpl60, and its subsequent cleavage to gpl20 were not

1799

Vol. 66, No. 3

1800 NOTES

gpl2Q gp41

-f~c N0I'NS NS N NT N'.11-*

HA HA H HV

5aP1)tf'0} CFCjt\;3'C + t- +-- t t

RcdtwICd 111fCCti\iti I)ZZ.

_ - ,J

H

Mock WT 611 616 624 637 674 518

p64 -

p53 -

gp4l -

V,gm: - 92.5

-MS *flg~ an a S c - 69- am - am am - ur

A- 46!-- _w -m wa _u ft4-o46

p34 - _ _ _ _3 0

p24 - _ _ ___w4ww4

7

Mock WT 611 616 624 637 674 518

WT 611 613 616 618 WT 624 637 639 674

-gpl2

FIG. 1. (Top) Locations of five potential N-linked glycosylationsites within the extracellular domain of HXB2 gp4l. C, conservedcysteine residue; closed circles, the five potential N-linked glyco-sylation sites (Asn-X-Ser/Thr); hatched box, the hydrophobic mem-brane anchor sequence; N, S, T, H, A, and V, asparagine, serine,threonine, histidine, alanine, and valine, respectively; numbers,amino acid positions according to the numbering system of Ratner etal. (17); "Slight," mutants with subtle impairment of viral infectiv-ity. (Bottom) Expression of gpl60 and gp120 in COS-7 cells trans-fected by the wild-type (WT) or mutant viruses. Cell lysatesprepared from COS-7 transfectants were reacted with a referenceHIV-1-positive human serum. The positions of gpl60 and gpl20 are

indicated.

drastically affected by the alteration of the canonicalN-linked glycosylation sequences at all four highly con-

served N-linked glycosylation sites and a fifth nonconservedN-linked glycosylation site within the extracellular domainof gp4l.

Lysates prepared from virions collected from the culturesupernatants of the COS-7 cultures transfected by the wild-type virus or the mutant viruses were also analyzed byWestern blot with a pool of HIV-1-positive sera showingpositive reactivity to gp4l. As shown in Fig. 2, alterationsintroduced at the four conserved N-linked glycosylationsites caused all four mutated gp4l species to migrate fasterthan the wild-type gp4l. In contrast, an identical substitutionof an asparagine residue with a histidine residue at thenonconserved site 674 had no discernible effect on themigration pattern of this mutated gp4l. The specificity ofgp4l was indicated by its reactivity to a rabbit anti-gp4l

TABLE 1. Synthetic oligonucleotides used for the constructionof HIV-1 mutants

Mutant Mutagenic oligonucleotide (5' to 3')a

518 GCAGTGGGAATAGaAGCTTaGTTCCTTGGGTTCTTG611 GTGCCTTGGcATGCTAGTTG613 CTTGGAATGCTgcTTGGAGTAATAA

616 TTGGAGTcATAAATCTCTGGAA

618 GGAGTAATAAAgCTtTGGAACAGAT624 CAGATCTGGcATCACACGACC637 AGAAATTAACcATTACACAAGCT639 AGAGAAATTAACAATTACgtcAGCTTAATACACTCCTT

674 GAATTGGTTTcACATAACAAATTG

a Lowercase letters indicate mutation sites.

9v:

gp4l - ^^

- 925- 69

46

A - 30

Mock WT 611 616 624 518 637 674

92 569

p64 - vow"mU "mmp53 - _m 4_ am se s - 46

gp4l - 40* 3 _ _p34- _ _e _

q.p4gM4 -30

p24 - e _b ___

FIG. 2. Detection of wild-type and mutant gp4l in cell-freevirions by Western blot analysis. Viral lysates were prepared fromcell-free supernatants of the wild-type (WT)- or mutant (611, 616,624, 637, 674, and 518)-transfected COS-7 cells. (Top) Viral lysateswere reacted with a pool of four HIV-1-positive human serumsamples. The positions ofpol gene-encoded p64, p53, and p34, gaggene-encoded p24, and gp4l are indicated. (Middle) Viral lysateswere reacted with a rabbit anti-gp4l serum. (Bottom) Viral lysateswere treated with PNGase-F (Genzyme Corp., Cambridge, Mass.)according to the manufacturer's suggested procedures. The treatedlysates were analyzed by Western blot with the same pool ofHIV-1-positive human serum mentioned in the legend to Fig. 1.gp4l*, position of the deglycosylated gp4l.

serum (kindly provided by S. Alexander of CambridgeBiotech Corp., Worcester, Mass.) and by the absence ofgp4l in the virions prepared from mock-transfected COS-7cells or from mutant 518-transfected COS-7 cells (Fig. 2).Mutant 518 had a stop codon inserted at the seventh aminoacid residue of gp4l (Table 1) and was not expected to havegp4l in its virions. One possible interpretation of thesefindings is that N-linked sugars are present at all four highlyconserved potential N-linked glycosylation sites. This inter-pretation is particularly attractive for mutants 611 and 616 inlight of our finding that after virions were treated withpeptide-N4-(N-acetyl-p-glucosaminyl)asparagine amidase, thedeglycosylated gp4l of mutants 611 and 616 comigrated withthe deglycosylated gp4l of the wild-type virus (Fig. 2).

J. VIROL.

NOTES 1801

Luuo Mock* WTAa 611* 613

100 a

3 6 10 13 17 20 24DAYS POST-INFECTION

100,

80

60

I-20

0L

s-100-CI.-cc

E

DAYS POST-INFECTION

80

60Er

40c2-

20

O0

DAYS POST-INFECTION DAYS POST-INFECTION

Mock

-* WT

a 674

° 80 /

40-

___

3 6 10 13 17 20 24DAYS POST-INFECTION

FIG. 3. In vitro growth properties of the wild-type and N-linked glycosylation mutants in CD4-positive SupTl cells. The procedure for theRT assay was described previously (26). "Mock," mock-infected cultures; WT, wild type.

To study whether HIV-1 infectivity was affected by alter-ations introduced to these five N-linked glycosylation sitesof gp4l, cell-free virions were collected 48 h posttransfectionfrom COS-7 cultures transfected with mutant and wild-typevirus. After adjustments for the level of reverse transcriptase(RT) activity were made, equal amounts of mutant andwild-type virus were used to infect a CD4-positive cell line,SupTl. The RT activities in the supernatants of the infectedSupTI cultures were monitored every 3 or 4 days. Amongthe five N-linked glycosylation site mutants, the growthkinetics of mutant 637 was substantially delayed comparedwith that of the wild-type virus (Fig. 3). While the RTactivity was detected in the wild-type-virus-infected cultures10 days postinfection, the RT activity was not detected inmutant 637-infected cultures until 27 days postinfection (Fig.3). Slight delays in growth kinetics as monitored by the RTactivity were observed in cultures infected by mutant 611 or

mutant 616, with mutant 611 showing more delay thanmutant 616. No delay in growth kinetics was observed formutant 624 or mutant 674. These results suggest that some ofthe conserved N-linked glycosylation sites within the extra-cellular domain of gp4l may play roles in HIV-1 infectivity.The infectivities of three other N-linked glycosylation site

mutants, designated 613, 618, and 639 (Fig. 1), which had the

serine or threonine residue of the canonical Asn-X-Ser/Thrsequence replaced by a noncanonical alanine or valineresidue, were also studied. These third-site mutants were

constructed to study whether the observed effects on viralinfectivity for mutants 611, 616, and 637 were due to aminoacid substitutions introduced at the first site of the canonicalAsn-X-Ser/Thr sequence per se rather than to the loss ofN-linked glycosylation sites. If a third-site mutant has a

phenotype similar to that of the wild-type virus, it is highlyunlikely that the observed effect on viral infectivity of thecorresponding first-site mutant can be attributed to the lossof an N-linked glycosylation site. Both gpl60 and gpl20 weredetected in COS-7 cells transfected by mutant 613 (Fig. 1).No delay in growth kinetics was observed for this mutantcompared with the wild-type virus (Fig. 3). Thus, the ob-served effect on viral infectivity for mutant 611 was likelydue to the substitution of a histidine residue for an aspara-gine residue.For mutant 618, no RT activity in cultures infected by this

mutant was detected throughout the entire follow-up period(Fig. 3). The relative amounts of gpl60 and gpl20 in thewild-type-virus-transfected COS-7 cells were compared byWestern blot analysis with those in the mutant 618-trans-fected COS-7 cells. As shown in Fig. 1, the amount of gpl20

Q

o Mock* WT

- a 624

o Mock* WTA 637* 639

VOL. 66, 1992

1802 NOTES

C ..vsa:es

MW Mlock W- 618200 --

gp 120925- ie *w 4

69-- a

MocK E9'8

b

46

FIG. 4. Radioimmunoprecipitation analysis of gpl20 present incell lysates and virus-free culture supernatants of mutant 618-transfected COS-7 cultures. Cell lysates and virus-depleted culturesupernatants were derived from COS-7 cultures metabolically la-belled with [35S]cysteine and transfected with no DNA (mock),wild-type DNA (WT), or mutant 618 proviral DNA and reacted witha reference HIV-1-positive human serum. The positions of gp160and gp120 are indicated. MW, molecular weight standards.

in the mutant 618-transfected COS-7 cells was disproportion-ate to that of gpl60 (Fig. 1). It appears that the phenotype ofmutant 618 can be explained by the excessive secretion ofgp120, possibly as the result of dissociation from gp4l. Asshown in Fig. 4, gpl60 and gpl20 in the wild-type-virus-transfected COS-7 cells were detected by radioimmunopre-cipitation analysis (11). In COS-7 cells transfected by mutant618, however, gpl60 was more readily detected than gpl20and most of the gpl20 was detected in the culture superna-tant. The loss of infectivity by mutant 618 could be attributedto excessive secretion of gpl20. It is unlikely that the loss ofinfectivity is caused by the loss of N-linked sugars, since asimilar phenotype was not observed with mutant 616. Thesefindings leave the question of whether N-linked sugars havea role in the slight impairment of infectivity by mutant 616unresolved. It should be noted, however, that the amino acidresidue altered by mutant 618 is outside the regions previ-ously reported to be critical for the association betweengpl20 and gp4l (9).

Severe impairment of infectivity was observed for mutant639. The level of RT activity detected in cultures infected bymutant 639 was not significantly above that of the mock-infected cultures until 20 days postinfection, which was adelay of 10 days compared with the wild-type virus (Fig. 3).Both gpl60 and gpl20 were detected in the mutant 639-transfected COS-7 cells (Fig. 1). The mutated gp4l speciesdetected in the virions also migrated faster than the wild-type virus (data not shown), which is compatible with thesuggestion that N-linked sugars may be present in thisconserved N-linked glycosylation site. One possible inter-pretation of the finding that both mutant 637 and mutant 639have severe impairment of infectivity is that N-linked sugarsat this particular N-linked glycosylation site may have im-portant roles in HIV-1 infectivity. However, the alternativeinterpretation that the impairment of viral infectivity wascaused by amino acid substitutions per se rather than by theloss of N-linked sugars cannot be ruled out.The N-linked sugars of HIV-1 have been proposed as

potential targets for antiviral therapeutic agents (6, 7, 12, 13,24). In fact, some inhibitors of the N-linked glycosylationpathway were previously shown to reduce HIV-1 infectivity(13, 16). However, it remains unclear which HIV-1 glyco-

proteins are targeted by these inhibitors. Our results raisethe possibility that glycosylation inhibitors aimed at the gp4lN-linked glycosylation site represented by mutants 637 and639 may have an antiviral effect.

We thank Y. Chow and Z. Matsuda for helpful discussions; A.Wolf, B. Du, and M. F. McLane for technical assistance; and E.Conway for editorial assistance.

This work was supported by Public Health Service grants CA-39805 and HL-33774 from the National Institutes of Health andcontract DAMD 17-90-C-0151 from the U.S. Army. W.-R. Lee wassupported by training grant D43TW00004 from the Fogarty Interna-tional Center, National Institutes of Health.

REFERENCES1. Allan, J., J. E. Coligan, F. Barin, M. F. McLane, J. G. Sodroski,

C. A. Rosen, W. A. Haseltine, T. H. Lee, and M. Essex. 1985.Major glycoprotein antigens that induce antibodies in AIDSpatients are encoded by HTLV-III. Science 228:1091-1094.

2. Berman, P. W., W. M. Nunes, and 0. K. Haffar. 1988. Expres-sion of membrane-associated and secreted variants of gp160 ofhuman immunodeficiency virus type 1 in vitro and in continuouscell lines. J. Virol. 62:3135-3142.

3. Chou, M. J., T. H. Lee, A. Hatzakis, T. Mandalaki, M. F.McLane, and M. Essex. 1988. Antibody responses in earlyhuman immunodeficiency virus type I infection in hemophiliacs.J. Infect. Dis. 157:805-811.

4. Fisher, A. G., E. Collalti, L. Ratner, R. C. Gallo, and F.Wong-Staal. 1985. A molecular clone of HTLV-III with biolog-ical activity. Nature (London) 316:262-265.

5. Gallaher, W. R., J. M. Ball, R. F. Garry, M. C. Griffin, andR. C. Montelaro. 1989. A general model for the transmembraneproteins of HIV and other retroviruses. AIDS Res. Hum.Retroviruses 5:431-440.

6. Gruters, R. A., J. J. Neetjes, M. Tersmette, R. E. de Goede, A.TuIp, H. G. Huisman, F. Miedema, and H. L. Ploegh. 1987.Interference with HIV-induced syncytiurn formation and viralinfectivity by inhibitors of trimming glucosidase. Nature (Lon-don) 330:74-77.

7. Karpas, A., G. W. Fleet, R. A. Dwek, S. Petursson, S. K.Namgoong, N. G. Ramsden, G. S. Jacob, and T. W. Radema-cher. 1988. Amino sugar derivatives as potential anti-humanimmunodeficiency virus agents. Proc. Natl. Acad. Sci. USA85:9229-9233.

8. Kowalski, M., L. Bergeron, T. Dorfman, W. Haseltine, and J.Sodroski. 1991. Attenuation of human immunodeficiency virustype 1 cytopathic effect by a mutation affecting the transmem-brane envelope glycoprotein. J. Virol. 65'281-291.

9. Kowalski, M., J. Potz, L. Basiripour, T. Dorfman, W. C. Goh, E.Terwilliger, A. Dayton, C. Rosen, W. Haseltine, and J. Sodroski.1987. Functional regions of the envelope glycoprotein of humanimmunodeficiency virus type 1. Science 237:1351-1355.

10. Kunkel, T. A., J. D. Roberts, and R. A. Zakour. 1987. Rapid andefficient site-specific mutagenesis without phenotypic selection.Methods Enzymol. 154:367-382.

11. Lee, T. H., J. E. Coligan, T. Homma, M. F. McLane, N.Tachibana, and M. Essex. 1984. Human T-cell leukemia virus-associated membrane antigens (HTLV-MA): identity of themajor antigens recognized following virus infection. Proc. Natl.Acad. Sci. USA 81:3856-3860.

12. Matthews, T. J., K. J. Weinhold, H. K. Lyerly, A. J. Langlois,H. Wigzell, and D. P. Bolognesi. 1987. Interaction between thehuman T-cell lymphotropic virus type IIIB envelope glycopro-tein gpl20 and the surface antigen CD4: role of carbohydrate inbinding and cell fusion. Proc. Natl. Acad. Sci. USA 84:5424-5428.

13. Montefiori, D. C., W. E. Robinson, Jr., and W. M. Mitchell.1988. Role of protein N-glycosylation in pathogenesis of humanimmunodeficiency virus type 1. Proc. Natl. Acad. Sci. USA85:9248-9252.

14. Muesing, M. A., D. H. Smith, C. D. Cabradilla, C. V. Benton,L. A. Lasky, and D. J. Capon. 1985. Nucleic acid structure andexpression of the human AIDS/lymphadenopathy retrovirus.

J. VIROL.

NOTES 1803

Nature (London) 313:450-458.15. Myers, G., A. B. Rabson, S. F. Josephs, T. F. Smith, and F.

Wong-Staal. 1989. Human retroviruses and AIDS 1989: a com-pilation and analysis for nucleic acid and amino acid sequences.Los Alamos National Laboratory, Los Alamos, N. Mex.

16. Pal, R., V. S. Kalyanaraman, G. M. Hoke, and M. G. Sarnga-dharan. 1989. Processing and secretion of envelope glycopro-teins of human immunodeficiency virus type 1 in the presence oftrimming glucosidase inhibitor deoxynojirimycin. Intervirology30:27-35.

17. Ratner, L., W. Haseltine, R. Patarca, K. J. Livak, B. Starcich,S. F. Josephs, E. R. Doran, J. A. Rafalski, E. A. Whitehorn, K.Baumeister, L. Ivanoff, S. R. Petteway, Jr., M. L. Pearson, J. A.Lautenberger, T. S. Papas, J. Ghrayeb, N. T. Chang, R. C.Gallo, and F. Wong-Staal. 1985. Complete nucleotide sequenceof the AIDS virus, HTLV-III. Nature (London) 313:277-284.

18. Sanchez Pescador, R., M. D. Power, P. J. Barr, K. S. Steimer,M. M. Stempien, S. L. Brown Shimer, W. W. Gee, A. Renard, A.Randolph, and J. A. Levy. 1985. Nucleotide sequence andexpression of an AIDS-associated retrovirus (ARV-2). Science227:484-492.

19. Sanger, F., and A. R. Coulson. 1975. A rapid method fordetermining sequences in DNA by primed synthesis with DNApolymerase. J. Mol. Biol. 94:441-445.

20. Syu, W. J., W. R. Lee, B. Du, Q. C. Yu, M. Essex, and T. H.Lee. 1991. Role of conserved gp4l cysteine residues in the

processing of human immunodeficiency virus envelope precur-sor and viral infectivity. J. Virol. 65:6349-6352.

21. VanBeveren, C., J. Coffin, and S. Hughes. 1985. Nucleotidesequences completed with functional and structural analysis,appendix B, p. 567-1148. In R. Weiss, N. Teich, H. Varmus,and J. Coffin (ed.), RNA tumor viruses, 2nd ed., vol. 2.Supplements and indexes. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

22. Veronese, F. D., A. L. DeVico, T. D. Copeland, S. Oroszlan,R. C. Gallo, and M. G. Sarngadharan. 1985. Characterization ofgp4l as the transmembrane protein coded by the HTLV-III/LAV envelope gene. Science 229:1402-1405.

23. Wain Hobson, S., P. Sonigo, 0. Danos, S. Cole, and M. Alizon.1985. Nucleotide sequence of the AIDS virus, LAV. Cell40:9-17.

24. Walker, B. D., M. Kowalski, W. C. Goh, K. Kozarsky, M.Krieger, C. Rosen, L. Rohrschneider, W. A. Haseltine, and J.Sodroski. 1987. Inhibition of human immunodeficiency virussyncytium formation and virus replication by castanospermine.Proc. Natl. Acad. Sci. USA 84:8120-8124.

25. Yu, X. F., S. Ito, M. Essex, and T. H. Lee. 1988. A naturallyimmunogenic virion-associated protein specific for HIV-2 andSIV. Nature (London) 335:262-265.

26. Yu, X. F., M. Matsuda, M. Essex, and T. H. Lee. 1990. Openreading frame vpr of simian immunodeficiency virus encodes avirion-associated protein. J. Virol. 64:5688-5693.

VOL. 66, 1992